Anna
Bernardi
a,
Jesus
Jiménez-Barbero
b,
Alessandro
Casnati
c,
Cristina
De Castro
d,
Tamis
Darbre
e,
Franck
Fieschi
f,
Jukka
Finne
g,
Horst
Funken
h,
Karl-Erich
Jaeger
h,
Martina
Lahmann
i,
Thisbe K.
Lindhorst
j,
Marco
Marradi
k,
Paul
Messner
l,
Antonio
Molinaro
d,
Paul V.
Murphy
m,
Cristina
Nativi
n,
Stefan
Oscarson
o,
Soledad
Penadés
k,
Francesco
Peri
p,
Roland J.
Pieters
q,
Olivier
Renaudet
r,
Jean-Louis
Reymond
e,
Barbara
Richichi
n,
Javier
Rojo
s,
Francesco
Sansone
c,
Christina
Schäffer
l,
W. Bruce
Turnbull
t,
Trinidad
Velasco-Torrijos
u,
Sébastien
Vidal
v,
Stéphane
Vincent
w,
Tom
Wennekes
x,
Han
Zuilhof
xy and
Anne
Imberty
*z
aUniversità di Milano, Dipartimento di Chimica Organica e Industriale and Centro di Eccellenza CISI, via Venezian 21, 20133 Milano, Italy
bCentro de Investigaciones Biológicas, CSIC, 28040 Madrid, Spain
cUniversità degli Studi di Parma, Dipartimento di Chimica, Parco Area delle Scienze 17/a, 43100 Parma, Italy
dDepartment of Chemical Sciences, Università di Napoli Federico II, Complesso Universitario Monte Santangelo, Via Cintia 4, I-80126 Napoli, Italy
eDepartment of Chemistry and Biochemistry, University of Berne, Freiestrasse 3, CH-3012, Berne, Switzerland
fInstitut de Biologie Structurale, 41 rue Jules Horowitz, 38027 Grenoble Cedex 1, France
gDepartment of Biosciences, University of Helsinki, P. O. Box 56, FI-00014 Helsinki, Finland
hInstitute of Molecular Enzyme Technology, Heinrich-Heine-University Düsseldorf, Forschungszentrum Jülich, D-42425 Jülich, Germany
iSchool of Chemistry, Bangor University, Deiniol Road Bangor, Gwynedd LL57 2UW, UK
jOtto Diels Institute of Organic Chemistry, Christiana Albertina University of Kiel, Otto-Hahn-Platz 3-4, D-24098 Kiel, Germany
kLaboratory of GlycoNanotechnology, CIC biomaGUNE and CIBER-BBN, P° de Miramón 182, 20009 San Sebastián, Spain
lDepartment of NanoBiotechnology, NanoGlycobiology Unit, University of Natural Resources and Life Sciences, Muthgasse 11, A-1190 Vienna, Austria
mSchool of Chemistry, National University of Ireland, Galway, University Road, Galway, Ireland
nDipartimento di Chimica, Universitá degli Studi di Firenze, Via della Lastruccia, 13, I-50019 Sesto Fiorentino – Firenze, Italy
oCentre for Synthesis and Chemical Biology, UCD School of Chemistry and Chemical Biology, University College Dublin, Belfield, Dublin 4, Ireland
pOrganic and Medicinal Chemistry, University of Milano-Bicocca, Piazza della Scienza, 2, 20126 Milano, Italy
qDepartment of Medicinal Chemistry and Chemical Biology, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, P.O. Box 80082, 3508 TB Utrecht, The Netherlands
rDépartement de Chimie Moléculaire, UMR-CNRS 5250 & ICMG FR 2607, Université Joseph Fourier, BP53, 38041 Grenoble Cedex 9, France
sGlycosystems Laboratory, Instituto de Investigaciones Químicas, CSIC – Universidad de Sevilla, Av. Américo Vespucio, 49, Seville 41092, Spain
tSchool of Chemistry and Astbury Centre for Structural Molecular Biology, University of Leeds, Leeds, LS2 9JT, UK
uDepartment of Chemistry, National University of Ireland, Maynooth, Maynooth Co. Kildare, Ireland
vInstitut de Chimie et Biochimie Moléculaires et Supramoléculaires UMR 5246, CNRS, Université Claude Bernard Lyon 1, 43 Boulevard du 11 Novembre 1918, F-69622 Villeurbanne, France
wUniversity of Namur (FUNDP), Département de Chimie, Laboratoire de Chimie Bio-Organique, rue de Bruxelles 61, B-5000 Namur, Belgium
xLaboratory of Organic Chemistry, Wageningen University, Dreijenplein 8, 6703 HB Wageningen, The Netherlands
yDepartment of Chemical and Materials Engineering, King Abdulaziz University, Jeddah, Saudi Arabia
zCentre de Recherche sur les Macromolécules Végétales (CERMAV – CNRS), affiliated with Grenoble-Université and ICMG, F-38041 Grenoble, France. E-mail: Imberty@cermav.cnrs.fr
First published on 19th December 2012
Multivalency plays a major role in biological processes and particularly in the relationship between pathogenic microorganisms and their host that involves protein–glycan recognition. These interactions occur during the first steps of infection, for specific recognition between host and bacteria, but also at different stages of the immune response. The search for high-affinity ligands for studying such interactions involves the combination of carbohydrate head groups with different scaffolds and linkers generating multivalent glycocompounds with controlled spatial and topology parameters. By interfering with pathogen adhesion, such glycocompounds including glycopolymers, glycoclusters, glycodendrimers and glyconanoparticles have the potential to improve or replace antibiotic treatments that are now subverted by resistance. Multivalent glycoconjugates have also been used for stimulating the innate and adaptive immune systems, for example with carbohydrate-based vaccines. Bacteria present on their surfaces natural multivalent glycoconjugates such as lipopolysaccharides and S-layers that can also be exploited or targeted in anti-infectious strategies.
Anne Imberty | Dr. Imberty is Research Director at the Centre de Recherches sur les Macromolécules Végétales (CERMAV), affiliated to the Centre National de la Recherche Scientifique (CNRS) based at Grenoble, France. She graduated in biology from Ecole Normale Supérieure in Paris. In 1984, she joined the CNRS in Grenoble and did her PhD on starch structure. She started studying protein-carbohydrate interaction during her post-doc in Toronto. Since 1999 she has a senior research position in CNRS-Grenoble and is the head of the Molecular Glycobiology group. Her research interests are in the field of structural glycobiology, with main interest on biologically active oligosaccharides and their interaction with lectins and glycosyltransferases. She now focuses on characterization of the molecular basis of recognition between lectins from pathogens and human glycoconjugates and design of glycocompounds with anti-infectious properties. |
Protein–carbohydrate interactions frequently mediate the first step of the infection process for many pathogens including viruses, fungi, bacteria, and bacterial toxins.26–29 Therefore, a vast array of unnatural glycoconjugates (neoglycoconjugates) with various valencies and spatial arrangement of the ligands have been constructed to prevent or treat diseases caused by pathogens. Scaffolds based on proteins,30–32 polymers,33–36 calixarenes,37–40 dendrimers,41–45 cyclodextrins,46,47 cyclopeptides,48–50 fullerenes,51,52 gold nanoparticles,53,54 and quantum dots55–57 provide nanoscale materials with anti-adhesive and cell targeting properties. Such structures that can competitively interfere with the recognition processes between host cells and pathogens have the potential to prevent colonisation or even reverse the formation of biofilms. Another alternative to fight pathogens relies on the utilization of glycoconjugates that can act as vaccines and immunomodulators. Vaccines have long relied on attenuated strains of microorganisms as a means of delivering the extracellular carbohydrate antigens. As the cell surface of pathogenic bacteria and viruses is often covered with unusual carbohydrates, structurally defined glycoconjugates displaying these structures are starting to emerge as the vaccines for the future.58,59 Following a lectin-mediated cellular uptake mechanism, such carbohydrate-based vaccines can prepare the immune defense mechanisms in advance of an infection, or to stimulate the body to protect itself against an existing chronic infection.
This review will give an overview of synthetic or natural multivalent glycoconjugates that can be used to inhibit the adhesion of viruses, bacterial toxins, and bacteria to host cells or to stimulate the innate and adaptive immune systems against these pathogens. In addition to the pioneering work of several groups mainly from North America,7,17,25,60,61 this field is currently flourishing in Europe, and this review is aimed at giving an overview of this very active domain.
Glycodendritic compounds have been used as tools to study and to interfere with infectious processes in which DC-SIGN is involved with the aim to develop new antiviral drugs and immune modulators.63 The conjugation of glycomimetics to dendritic compounds has provided multivalent compounds with interesting antiviral activity. IC50s in the low nanomolar range have been obtained in biological assays using pseudotyped Ebola viral particles.64 Also, these compounds present very good activity as inhibitors of HIV trans-infection of T-cells, a more relevant infection model where DC-SIGN is implicated.65,66
Glycodendritic structures have also been prepared recently using a convergent approach with extensive use of the Cu(I) catalyzed azide–alkyne cycloaddition (CuAAC) reaction also known as “click” chemistry (Fig. 1). In this new strategy, first and second generation glycodendrons have been prepared conveniently functionalized for further conjugation on different scaffolds including a fullerene molecule or a virus-like particle protein.52,67
Fig. 1 (a) Glycodendrofullerene 1 with 36 mannoses; (b) glycodendrimer 2 with 18 mannoses prepared using a CuAAC click reaction. |
The new glycodendritic compounds display a wide variety of valencies and spatial presentation of carbohydrate ligands. The glycodendrofullerenes (e.g., 1) prepared using this strategy are soluble under physiological conditions and present a very low cellular toxicity. The globular disposition of carbohydrates on this spherical scaffold provides an interesting multivalent system which allows the carbohydrates to be recognized by lectins in a multivalent manner. Antiviral activity of these compounds using pseudotyped Ebola viral particles is in the micromolar range.52,67
In another attempt to mimic the cluster presentation of high-mannose-type glycans on the HIV envelope, gold nanoparticles biofunctionalized with oligomannosides (manno-GNPs, 3a–3d, Fig. 2) of gp120 high-mannose type glycans have been prepared and tested as anti-HIV agents. These manno-GNPs inhibited the DC-SIGN/gp120 binding in the micro- to nanomolar range, while the corresponding monovalent oligomannosides required millimolar concentrations, as measured by surface plasmon resonance (SPR) experiments.68 Furthermore, manno-GNPs were able to inhibit the DC-SIGN-mediated HIV trans-infection of human activated peripheral blood mononuclear cells at nanomolar concentrations in an experimental setting, which mimics the natural route of virus transmission from dendritic cells to T lymphocytes.69
Fig. 2 Gold nanoparticles 3a–3d bearing high-mannose type glycans (manno-GNPs) present on HIV envelope glycoprotein gp120 as anti-HIV synthetic glycoconjugates. |
Structurally defined bivalent lactose-containing clusters have been designed for optimal binding to galectins.41,72–74 These compounds were evaluated for binding to the entire set of adhesion/growth-regulatory galectins from chicken. Differential sensitivities were detected between distinct galectin forms within the chicken series. Two of the bivalent glycoclusters, 4 and 5 (Fig. 3), were identified as sensors for different galectin subtypes. Most pronounced were the selectivities of these two glycoclusters for the chimera-type galectin (galectin-3).
Fig. 3 Bivalent glycoclusters 4 (acyclic) and 5 (macrocyclic), identified as sensors for different galectin subtypes. |
The advent of ‘click’ chemistry in combination with chemoenzymatic synthesis of the complex oligosaccharide enabled the assembly of multivalent versions of the GM1os77 into glycodendrimers (Fig. 4).78 Divalent compound 7a was almost 10000 times more potent as an inhibitor of CTB binding to GM1os than monovalent compound 6. This enhancement in inhibitory potency is related to the multivalent interactions between CTB and the divalent compound 7a, as the latter does not show an enhancement in binding to antibodies that do not allow multiple interactions.79
Fig. 4 Structures of GM1os- (7a, 8a and 9a) and galactose (7b, 8b and 9b)-based inhibitors of cholera toxin binding. |
An additional increase was observed for tetravalent 8a (83000 fold). The most complex glycodendrimer in this study, octavalent 9a, was 380000 fold more potent (IC50 = 50 pM and relative inhibitory potency of 47500 per GM1os).78 A detailed study of the mode of action revealed that complex aggregates between the inhibitor and toxin are formed. These are possible because of the mismatch between the valencies of the toxin (five) versus those of the inhibitors (two, four, eight).80
The galactose dendrimers 7b, 8b and 9b are a simplified glycomimetic version of the multivalent GM1 derivatives. The inhibitory potency did suffer due to this modification since the relatively large binding site of the B-subunit remains partly unoccupied resulting in a lower binding affinity. Nevertheless, multivalency effects were able to counteract the lower binding affinity of galactose and the inhibitory potencies of compounds 8b and 9b were shown to be competitive with the natural GM1os ligand.81
A series of ganglioside mimics, in which the non-interacting oligosaccharide backbone of GM1os was replaced by an appropriate cyclohexanediol, was chosen to reproduce the topological features of the 3,4-disubstituted galactose residue (Gal-II) in GM1os.82,83 The divalent presentation of a structurally simplified second generation mimic 10 on a functionalized calix[4]arene scaffold (Fig. 5) led to a 3800-fold (1900-fold per sugar mimic) enhancement of CTB affinity, thus reaching the potency of GM1os itself.84 Although computational studies show that the divalent ligand 11 could easily span two binding sites on cholera toxin, NMR data indicate that the action of this divalent ligand is likely to involve additional interactions between the linker and the protein.
Fig. 5 Structurally simplified GM1os mimic 10 grafted onto a functionalized calix[4]arene scaffold to give divalent ligand 11. |
Nevertheless, avidity effects have been frequently observed with a variety of multivalent mannose-containing glycomimetics, like 12–14 (Fig. 6a–c). Such avidity can originate from statistical effects arising from (i) a higher concentration of mannose in the proximity of the carbohydrate binding site, (ii) existence of additional carbohydrate binding sites on the lectin FimH, or (iii) occurrence of the natural multivalent process, since fimbriae occur on the bacterial surface in several hundreds of copies. Thus, mannose-terminated multivalent glycocompounds have become important to test mannose-specific bacterial adhesion in a supramolecular context.88,91 Meanwhile, testing of type 1 fimbriae-mediated bacterial adhesion has been greatly facilitated by employing GFP-transfected strains.92 Interestingly, adhesion on multivalent glycomaterials can be utilized for aggregating E. coli and removing them from solution with the use of appropriate filters. Glyconanodiamonds decorated with mannose (Fig. 6d) have been shown to be able to clean bacteria-polluted water.93
Fig. 6 Examples of various multivalent glycoconjugates inhibiting type 1 fimbriae-mediated bacterial adhesion. (a) Octopus glycosides 12; (b) glycodendrimer 13; (c) bifunctional ligand 14 to test multiple binding sites on FimH; (d) glyconanodiamonds to remove pathogenic bacteria from polluted water sources, a sandwich assay is displayed, utilizing two different bacterial strains. |
Fullerene hexakis-adducts bearing 12 peripheral mannose moieties (15–20) have been prepared by grafting sugar derivatives onto the fullerene core52 and assayed as inhibitors of FimH (Fig. 7).88 Dissociation constants in the range of 12 to 95 nM were measured using isothermal titration calorimetry (ITC), surface plasmon resonance (SPR) and hemagglutination assays. Most importantly, the number of possible interactions between the multimers and the lectin and the average binding strength per functional mannose unit could be measured. Thus, this study demonstrated for the first time that a globular C60 structure can accommodate up to seven FimH molecules.
Fig. 7 Dodecavalent mannofullerenes 15–20 as FimH inhibitors. |
Synthetic mono- and multivalent galabiose derivatives 21a–d (Fig. 8) inhibited bacterial adhesion to the coated chip surfaces in a dose-dependent manner (Table 1). An octavalent galabiose compound 21d was superior to the tetravalent derivative 21c, which in turn was a better inhibitor than the monovalent galabiose derivative 21a. However, the multivalency effect was much more pronounced in the case of the Streptococcus suis adhesion when compared to E. coli PapG. On the other hand, a more significant multivalency effect was observed in the inhibition of the mannose-specific type-1-fimbriated E. coli with similar multivalent mannose molecules.97 It would appear that multivalent inhibitors do not reach multiple E. coli adhesin molecules as effectively as in the case of other bacteria such as S. suis (Table 1), and therefore the spacing of the binding sites in the adhesins may differ.
Fig. 8 Oligovalent galabiose derivatives. |
LecA (also called PA-IL) is a tetrameric cytotoxic lectin consisting of four subunits of 121 amino acids (12.75 kDa)98,102 with specificity for α-D-galactose and binding preferentially to Galα(1-4)Gal containing globotriaosylceramide Gb3 sphingolipid. The LecA crystal structure demonstrated the structural basis of the affinity for galactose monosaccharides with the participation of a calcium ion in the binding site.103 In addition to its cytotoxicity, it has been suggested that this lectin contributes to the formation of bacterial microcolonies and the formation of biofilms.104
Many of multivalent glycoconjugates have been synthesized for inhibiting the binding of LecA to galactosylated surfaces (Table 2 and Fig. 9). High-valency compounds such as galactosylated helical poly(phenylacetylene) polymer 29,34 fullerenes 23,51 glyconanoparticles 31,105 or glycodendrimers 32106 are efficient ligands for inhibition, but their aggregative properties and the strong resulting precipitation create difficulties for measuring affinity constants. Excellent results were obtained with calix[4]arenes 28,38,107 calix[6]arenes 26, β-peptoids 25, porphyrins 27108 and resorcin[4]arenes 22.109 Among these molecules, the 1,3-alternate conformer of calix[4]arene demonstrated the most efficient and dramatic increase in affinity. A chelate-binding mode with two galactose residues interacting with two neighbouring binding sites in a single LecA tetramer could be confirmed by the observation of well-defined nanometric fibers of lectin–glycocluster complexes through atomic force microscopy (AFM) study.110
Comp. | Valency | HIA (MIC) | ELLA (IC50) | SPR (Kd) | ITC (Kd) | Ref. |
---|---|---|---|---|---|---|
a β value calculated with galactose as reference. | ||||||
αMeGal | 1 | 150 μM | 50 μM | 106 | ||
βMeGal | 1 | 190 μM | 94 μM | 106 | ||
22 | 4 | Haemolysis | 0.7 μM, β = 315 | Not soluble | —/— | 109 |
23 | 12 | 0.78 μM, β = 12820 | 0.040 μM, β = 458 | 0.367 μM, β = 173 | —/— | 51 |
24 | 4 | > 2000 μM | —/— | 3.5 μM, β = 18 | 1.8 μM, β = 83 | 108 |
25 | 4 | > 2000 μM | —/— | 2.5 μM, β = 25 | 0.3 μM, β = 500 | 108 |
26 | 6 | 63 μM, β = 159 | —/— | 0.8 μM, β = 80 | 0.14 μM, β = 1071 | 108 |
27 | 4 | 63 μM, β = 159 | —/— | 1.4 μM, β = 45 | 0.33 μM, β = 454 | 108 |
28 | 4 | 500 μM, β = 20 | —/— | 0.5 μM, β = 144 | 0.176 μM, β = 852 | 38 |
29 | Polymer | 9 μM | —/— | —/— | 4.12 μM | 34 |
30 | 2 | —/— | 0.22 μM, β = 545 | —/— | —/— | 111 |
31 | 67 | 0.45 μM, β = 100 | —/— | 33 μM | 0.05 μM, β = 2824 | 105 |
32 | 9 | —/— | —/— | —/— | 230 nM, β = 409 | 106 |
33 | 4 | 0.78 μM, β = 4000a | —/— | —/— | 0.1 μM, β = 875 | 112 |
Fig. 9 Multivalent glycoconjugates 22–33 as LecA high affinity ligands. |
A recent study demonstrated that a divalent ligand 30 with an appropriate linker was sufficient to induce a chelation effect with LecA.111 A rigid spacer was designed based on the alternation of glucose moieties linked at the 1 and 4 positions by a 1,2,3-triazole unit via “click” chemistry. The resulting spacer was relatively rigid and straight and was linked to galactose units at both ends. The number of building blocks was varied, as well as the linker between the spacer and galactose ligand. This linker was shown to be of great importance and only compound 30 (Fig. 9) has the appropriate linker-length for achieving an inhibitory potency increase of 545-fold over a relevant reference compound.
Potent ligands for lectin LecA have been also obtained by synthesis of glycopeptide dendrimers GalAG2 33 and GalBG2 34 (Fig. 10).112 Multivalency strongly influenced binding, with the monovalent and divalent analogs showing much weaker interactions with the lectin. The much stronger binding of the phenyl galactosides to LecA compared to the thiopropyl-galactosides was explained by crystallographic analysis of the lectin–glycopeptide complexes, which revealed a specific interaction between a histidine residue on the lectin and the phenyl group in the ligands, while the thiopropyl side-chain was more disordered.112
Fig. 10 Structure of glycopeptide dendrimer inhibitors of P. aeruginosa biofilms. |
LecB (also called PA-IIL) is a tetramer consisting of four 11.73 kDa subunits with high specificity for L-fucose and a weaker one for D-mannose.98,102 The LecB crystal structure revealed the occurrence of two bridging calcium ions in the binding site. This unique mode of binding is not observed in other lectins,113 but explains the very high affinity for fucosides and Lewis a. Although most of the LecB is cytoplasmic, it could also be detected in the outer membrane, including on the surface of biofilm cells, from which it can be released by application of L-fucose.114 It has recently been hypothesized that LecB undergoes transient N-glycosylation that could play a role in the secretion mechanism.115 The search for a putative binding partner led to the proposal of outer membrane protein OprF which is a nonspecific, weakly cation-selective porin channel protein. LecB may mediate the adhesion of P. aeruginosa cells to receptors that are located on its surface and facilitate biofilm formation, thereby promoting colonization of host tissues.
The search for high affinity ligands for LecB initiated the synthesis of several classes of fucose-containing compounds (Table 3 and Fig. 11) based on calixarene 35,116 pentaerythritol 39117 or peptide dendrimer 36–38 scaffolds. Compounds have also been designed for bivalent presentation of αFuc(1 → 4)GlcNAc 40,118 and N-fucosyl amides 41.119
Fig. 11 Multivalent glycoconjugates 35–41 as LecB high affinity ligands. |
Glycopeptide dendrimer ligands for LecB were identified by screening combinatorial libraries of peptide dendrimers120–122 functionalized with N-terminal C-fucoside residues at the end of the dendrimer branches. FD2 36 and PA8 37 (Fig. 11) turned out to be potent ligands for LecB.123,124
Structure–activity relationship studies showed that multivalency was important for activity, in particular divalent and linear peptide analogs of the dendrimers showed strongly reduced binding at the level of monosaccharides (Table 4). These studies led to the identification of dendrimer 2G3 38 with 8 fucosyl endgroups as the most potent glycopeptide dendrimer ligand to LecB. The diastereoisomer D-36 prepared from D-amino acids was also demonstrated to be a similarly potent ligand to LecB.125
No | Structurea | Lectin | IC50, μM (ELLA)b | K D, μM (ITC)c | r.p./nd | Biofilm inhibitione |
---|---|---|---|---|---|---|
a Standard peptide notation with N-terminus at left and C-terminus at right. Amino acids are given in one-letter codes, italics indicate branching diaminoacids, B is L-2,3-diaminopropionic acid, the C-terminus (at right) is carboxamide (CONH2) in all cases. See also Fig. 10 for exemplification of the topology and the structure of the glycoside groups C-Fuc, GalA and GalB. b Enzyme-linked lectin assay. c Isothermal titration calorimetry. d r.p./n is the relative potency compared to the free sugar (L-fucose or D-galactose) per glycoside group. e Biofilm inhibition measured with the steel coupon assay at 50 μM. n.d. = not determined, — = no inhibition, + = less than 20% inhibition, ++ = 30 to 50% inhibition, +++ = up to 90% inhibition, ++++ = 100% inhibition. | ||||||
α-NPF | α-(p-Nitrophenyl)-L-fucoside | LecB | 5.27 ± 0.55 | 2.1 | — | |
T1 | (cFuc-RL)2BRIFV | LecB | 5 ± 0.45 | 1.7 | n.d. | |
KT1 37 | (cFuc)4(KRL)2BRIFV | LecB | 0.59 ± 0.059 | 7.2 | n.d. | |
2G0 | cFuc-KPL | LecB | 5.94 ± 1.24 | 1.9 | n.d. | |
2G1 | (cFuc-KPL)2KFKI | LecB | 2.7 ± 0.56 | 2.0 | n.d. | |
37 | (cFuc-KPL)4(KFKI)2KHI | LecB | 0.14 ± 0.035 | 20 | ++++ | |
Dd-36 | (cFuc-kpl)4(Kfkl)2Khl | LecB | 0.66 ± 0.12 | 4.2 | +++ | |
2G3 | (cFuc-KP)8(KLF)4(KKI)2KHI | LecB | 0.025 ± 0.005 | 55 | n.d. | |
NPG | p-Nitrophenyl β-D-galactopyranoside | LecA | 14.1 ± 0.2 | 6.2 | — | |
GalAG0 | GalA-KPL | LecA | 4.2 ± 0.9 | 21 | + | |
GalAG1 | (GalA-KPL)2KFKI | LecA | 0.5 ± 0.2 | 91 | ++ | |
GalAG2 33 | (GalA-KPL)4(KFKI)2KHI | LecA | 0.1 ± 0.01 | 220 | +++ | |
GalBG0 | GalB-KPL | LecA | 51.5 ± 6.7 | 1.7 | + | |
GalBG1 | (GalB-KPL)2KFKI | LecA | 2.1 ± 1.0 | 21 | ++ | |
GalBG2 34 | (GalB-KPL)4(KFKI)2KHI | LecA | 0.4 ± 0.1 | 60 | +++ |
BC2L-A has a strong affinity for α-D-mannosides (Kd of 2 μM for methyl α-D-mannopyranoside) and mannobioses. Bridging interaction with the branched trimannoside Manα1-3(Manα1-6)Man resulted in the formation of molecular strings as detected by protein crystallography and AFM. Oligomannose analogs presenting two mannosides separated by either rigid (42) or flexible (43) spacer arms were also tested (Fig. 12). Only the rigid linker yielded high affinity with a fast kinetics of clustering, while the flexible analog and the trimannoside displayed moderate affinities and no clustering.128
Fig. 12 Divalent mannosylated compounds as ligands of BC2L. |
Micelles formed from mannosylated poly(ethylene oxide)-b-poly(ε-caprolactone) (PEO-b-PCL) diblock copolymer and nanoparticles of poly(D,L-lactic acid) functionalized with the same copolymer have also been demonstrated to bind efficiently to BC2L-A.36,129
Synthetic derivatives of the receptor disaccharide galabiose on one hand, or polyvalent dendrimers of galabiose on the other hand (Fig. 8), have turned out to be exceptionally efficient inhibitors of S. suis adhesion, both at nanomolar concentrations.132,133 Inhibitors of the adhesion of S. suis to cells have also been found in natural sources, berries and juices, but their chemical nature has not yet been identified.134 In the characterization of the adhesion specificities and comparison of various mono- or oligovalent inhibitors, a live-bacteria application of surface plasmon resonance has turned out to be very useful.96
Due to phase variation, the expression of bacterial adhesins is not uniform even within a single bacterial strain. Therefore it may become important to be able to detect specifically bacteria that express a specific adhesin of potential clinical impact. Magnetic glyconanoparticles may represent ideal tools for this purpose. Magnetic beads containing derivatives of galabiose (Fig. 13) were capable of selecting bacteria in a mixture and determining their amounts in a luminescence assay.135 Optimisation of the carbohydrate ligand and its multivalent presentation in appropriate carriers are predicted to further improve the efficiency of the ligand in bacterial adhesion inhibition and detection.
Fig. 13 Use of galabiose-functionalized magnetic beads for identifying and isolating S. suis bacteria. |
LPSs isolated or extracted from their natural environment aggregate in solution due to their amphiphilic nature, resulting in a supramolecular aggregate which still displays multivalent features that can be of direct use. Indeed, it has recently been shown that the HIV protective antibody 2G12 is able to recognize the LPS produced from Rhizobium radiobacter Rv3.137 This particular LPS possesses a carbohydrate structure (Fig. 14) which resembles the epitope on the surface of HIV for which the antibody is specific. It has been observed that 2G12 interacted with lipooligosaccharides (LOS) only when the lipid moiety was present, so as to allow formation of a supramolecular aggregate in water (C. De Castro, personal communication).
Fig. 14 Structure of the LPS from Rhizobium radiobacter Rv3. The area in the box displays strong structural similarity with the epitope recognized from the mAb 2G12. LA stands for lipid A. |
Another interesting example of LPS multivalency relates to the stabilizing effect on the protein conformation embedded in the membrane. This concept guided the formulation of the new generation vaccine for Neisseria meningitidis group B (MenB),138 which is the causative agent of meningitis. The strain belonging to group B is the only member within its species for which the synthesis of a synthetic glycoprotein vaccine has not been successful so far as its capsular polysaccharide is not immunogenic. The current MenB vaccine is a vesicle prepared from fragments of the bacterial membrane; the antigenicity mainly arises from a selected pool of Neisseria proteins embedded in this artificial membrane, in which LPSs are thought to work as a multivalent glycoconjugate that stabilises the vesicle.
Given the high biological importance of LPSs as natural multivalent glycoconjugates in the elicitation/suppression of eukaryotic immunity, the structural and supramolecular study of such molecules plays an important role. In an attempt to create a “non-natural” LPS multivalent surface, these molecules have been extracted from different microbial sources and reconstituted in liposomes. The physico-chemical investigation of these systems has been performed by a combined experimental strategy, which has allowed characterisation at different observation scales, from the morphological to the micro-structural level.139,140 The next step will be the study of the elicitation of eukaryotic immunity of such an “artificial” bacterial surface.
Fig. 15 Multivalent glycosylated fullerenes for inhibition of LPS heptosyltransferase WaaC. |
The S-layer system is being exploited for multivalent glycan display based on the groundbreaking demonstrations that proteins can be recombinantly equipped with tailor-made glycosylation in an easily tractable bacterial system such as Escherichia coli143 and that glycosylation modules from different bacterial sources, including glycoproteins, lipopolysaccharides or exopolysaccharides, can be combined to achieve functional glycosylation.144 Using a combination of protein- and glycosylation-engineering approaches to produce self-assembling S-layer neoglycoconjugates, the feasibility of this system could be proven in vitro as well as in vivo, with the latter approach presenting interesting possibilities for live glycoconjugate delivery in future antipathogenic therapy.145
In a proof-of-concept study, the S-layer protein SgsE from Geobacillus stearothermophilus NRS 2004/3a (AF328862) was used as a matrix for the display of a branched heptasaccharide from the Campylobacter jejuni protein AcrA as well as for the E. coli O7 antigen.146 SgsE is a 903-amino acid protein which aligns in a 2D lattice with oblique (p2) symmetry and which is naturally O-glycosylated at multiple sites. The SgsE protein was engineered by including the signal peptide of PelB (pectate lyase from Erwinia carotovora) for periplasmic targeting. Furthermore, one of the natural protein O-glycosylation sites was engineered into an N-glycosylation site to be recognized by the heterologous oligosaccharyltransferase PglB. In this way, S-layer neo-glycoproteins could be produced based on plasmid-encoded glycosylation information for either of the model glycan structures. The degree of glycosylation of the S-layer neoglycoproteins after purification from the periplasmic fraction of the E. coli cell factory reached up to 100%. Electron microscopy revealed that recombinant glycosylation is fully compatible with the S-layer protein self-assembly system (Fig. 16). Thus, the S-layer system is a promising strategy for multivalent glycan display approaches, where strict (“nanometer-scale”) control over position and orientation of the glycan epitopes is desired.
Fig. 16 Model of a self-assembled SgsE-neoglycoprotein monolayer periodically displaying recombinant E. coli O7 antigens with nanometer-scale precision. Image reconstruction using Cinema 4 is based on a negatively stained preparation of the S-layer protein self-assembled in solution and on the pdb data of the glycans generated with Sweet at http://www.glycosciences.de/ (adapted from ref. 146, Wiley-VCH Weinheim). |
Fig. 17 Synthetic lipids with TLR4-modulating activity. |
TLR4 trigger can be remarkably sensitive and robust, stimulating prompt and powerful host defence responses to different species of invading bacteria. However, an excessively potent host response generates life-threatening syndromes such as acute sepsis and septic shock. Non-toxic LPSs or lipid A obtained from non-pathogenic bacteria such as Rhodobacter capsulatus and Rhodobacter sphaeroides are potent LPS antagonists in vitro148 although no molecules usable for pharmacological treatment for sepsis are not available yet.
Several synthetic molecules capable of modulating TLR4 activity have been developed.149,150 IAXO compounds 47–52, which include lipidated monosaccharides with an amino group on the C-6 position of the pyranose ring of D-glucose, were active in blocking the TLR4 signal in cells and in vivo models (Fig. 17).151 Their antagonist effect was due to the capacity of these molecules to displace from CD14 and TLR4-MD-2 complexes.152 These molecules are now commercially available with the proprietary name of IAXO (Innaxon) as selective small-molecule inhibitors of the TLR4 signal pathway, and lead compound IAXO-102 48 is in a preclinical phase as an antisepsis agent. More recently, lipid A analogues such as 53 were developed with a structure composed of two glucoside units connected by a linker both units bearing on C-4 an anionic sulfate group (Fig. 17).153 These compounds have an antagonist effect if administered together with endotoxin, and a mild agonist effect if administered alone. In the context of vaccine adjuvants acting on innate immunity receptors, natural LPS immobilized on nanoparticles was also demonstrated to have interesting TLR4 activating properties.154
Together with new studies on the molecular basis of the interactions between glycans and anti-HIV antibodies, gold nanoparticles functionalized with oligomannosides (manno-GNPs, Fig. 2) can offer an alternative in this direction. In order to gain deeper insights into the interactions between 2G12 and selected oligomannosides at the molecular level, the structural and affinity details of the 2G12/oligomannosides interactions have been studied by saturation transfer difference NMR spectroscopy (STD-NMR) and transferred NOE in isotropic solution.163 It was found that linear oligomannosides show a single binding mode to 2G12, with the non-reducing terminal disaccharide Man(α1-2)Man(α1-, making the closest antibody/oligosaccharide contacts in the bound state. In contrast, a branched pentamannoside showed two alternate binding modes involving both ligand arms, contrary to previous X-ray studies.164 Among the analysed series of ligands, the strongest 2G12 binders were the linear tri- and tetramannosides. This information is of key importance for the design of synthetic multivalent gp120 high-mannose mimics for HIV vaccine development. Indeed manno-GNPs were able to bind with high affinity and to interfere with the 2G12/gp120 binding as determined by SPR-based biosensors and STD-NMR.165 Cellular neutralization assays with manno-GNPs also demonstrated that GNPs coated with a linear tetramannoside could block the 2G12-mediated neutralization of a replication-competent HIV-1 under conditions that resemble those in which normal serum prevents infection of the target cell. All these results prove that selected manno-GNPs could function as an anti-adhesive barrier at an early stage of HIV infection, but also as synthetic mimics of the 2G12 epitope in the route of a carbohydrate-based vaccine against HIV.
Fig. 18 Gold nanoparticles bearing Galβ(1-4)Glcβ(1-6)[Galβ(1-4)]GlcNAcβ(1- (repeating units of S. pneumoniae type 14 capsular polysaccharide), T-helper ovalbumin (OVA) peptide OVA323–239 and D-glucose as carriers for synthetic carbohydrate-based vaccines. |
Footnote |
† Part of the Multivalent Scaffolds in Glycosciences themed issue. |
This journal is © The Royal Society of Chemistry 2013 |